Mixed-flow pump stability evaluation method based on sound and vibration fusion characteristic value

Abstract

The invention discloses a mixed-flow pump stability evaluation method based on sound and vibration fusion characteristic values. The method specifically comprises the following steps that 1, a spindle vibration displacement signal and a noise signal are collected; step 2, converting the spindle vibration displacement signal acquired in the step 1 into a time-frequency graph px by using continuous wavelet transform; step 3, converting the noise signals acquired in the step 1 into a time-frequency graph py by using continuous wavelet transform; step 4, solving average values Cx and Cy of similar numbers of the sliding windows and the fixed windows in the time-frequency graph px and the time-frequency graph py; and step 5, constructing Cx, y by combining Cx and Cy under different signals and different embedding dimensions, and calculating a fusion feature value Sx, y of the comprehensive spindle vibration displacement signal and the noise signal. The mixed-flow pump stability evaluation method based on the sound-vibration fusion characteristic value can quantify the stability of the mixed-flow pump during operation under a complex working condition by performing conjoint analysis on the data of the vibration measuring point and the noise measuring point of the main shaft of the mixed-flow pump.

Description

Technical Field

[0001] The invention belongs to the technical field of petrochemical industry and relates to a mixed flow pump stability evaluation method based on acoustic-vibration fusion characteristic values. Background Art

[0002] In the fields of petrochemicals, mixed flow pumps are widely used due to their excellent performance. When operating under stable conditions, mixed flow pumps run smoothly and with low noise. However, in complex conditions such as startup and speed change, the pump body vibrates violently and makes strong noise due to deteriorated stability, and even causes damage to mixed flow pump components and pipeline rupture, resulting in leakage of petrochemical raw materials. It can be seen that the operating stability of mixed flow pumps will directly affect the performance of petrochemical transportation systems in vibration and noise control and overall safety and stability. Traditional stability evaluation methods for mixed flow pumps mainly focus on signal analysis of a single measuring point, such as time domain, frequency domain and time-frequency analysis of noise signals. However, these methods are only applicable to evaluating the stability of its operation under simple conditions, but in the face of complex and changeable conditions, the information reflected by traditional stability evaluation methods is relatively one-sided, and it is impossible to accurately judge the impact of these complex vibrations and noises on the stability of mixed flow pumps. Summary of the invention

[0003] The purpose of the present invention is to provide a mixed flow pump stability evaluation method based on acoustic-vibration fusion eigenvalues. The method jointly analyzes the data of the mixed flow pump main shaft vibration measurement points and the noise measurement points. The mixed flow pump stability evaluation method based on acoustic-vibration fusion eigenvalues ​​can quantify the stability of the mixed flow pump when it is operating under complex working conditions.

[0004] The technical solution adopted by the present invention is a mixed flow pump stability assessment method based on acoustic-vibration fusion characteristic values, which specifically includes the following steps:

[0005] Step 1, collecting the spindle vibration displacement signal and noise signal;

[0006] Step 2: Use continuous wavelet transform to convert the spindle vibration displacement signal collected in step 1 into a time-frequency graph p x ;

[0007] Step 3: Use continuous wavelet transform to convert the noise signal collected in step 1 into a time-frequency graph p y ;

[0008] Step 4: Obtain the time-frequency graph p x And the time-frequency diagram p y The average value of the number of similarities between the sliding window and the fixed window C x and C y ;

[0009] Step 5: By combining different signals and C under different embedding dimensions xand C y Build C x,y , calculate the fusion eigenvalue S of the integrated spindle vibration displacement signal and noise signal x,y .

[0010] The present invention is also characterized in that:

[0011] The specific process of step 2 is:

[0012] Step 2.1, use the following formula (1) to calculate the scale parameter of the continuous wavelet transform of the vibration signal x(t):

[0013] a x =2·F xc ·T (1)

[0014] Among them, a x is the scale parameter; F xc is the center frequency of the wavelet function; T is the total scale of the continuous wavelet transform of the vibration signal x(t), which is calculated using the following formula (2):

[0015]

[0016] In the formula, f s is the sampling frequency of the data acquisition card of this experimental device;

[0017] Step 2.2: Use the following formula (3) to convert the scale parameter a x Convert to actual frequency:

[0018]

[0019] In the formula, F x is the actual frequency of the continuous wavelet transform of the vibration signal x(t);

[0020] Step 2.3, the continuous wavelet coefficients of the vibration signal x(t) are obtained by continuous wavelet transform through equation (4):

[0021]

[0022] Among them, K x is the continuous wavelet coefficient of the continuous wavelet transform of the vibration signal x(t); b is the translation parameter, is the wavelet function, and the complex Morlet wavelet function is used. The formula is as follows:

[0023]

[0024] Among them, F b is the bandwidth parameter; π is the circumference of a circle; e is a natural constant; i is an imaginary unit;

[0025] Step 2.4, using the actual frequency F x , continuous wavelet coefficient K x and sampling time t to draw the time-frequency diagram of the vibration displacement signal p x .

[0026] The specific process of step 3 is:

[0027] Step 3.1, use the following formula (6) to calculate the scale parameter of the continuous wavelet transform of the noise signal y(t):

[0028] a y =2·F yc ·T (6)

[0029] Among them, a y is the scale parameter; F yc is the center frequency of the wavelet function; T is the total scale of the noise signal y(t) for continuous wavelet transform, and the calculation formula is:

[0030]

[0031] In the formula, f s The sampling frequency of the data acquisition card of this experimental device;

[0032] Step 3.2: Use the following formula (8) to convert the scale parameter a y Convert to actual frequency:

[0033]

[0034] In the formula, F y is the actual frequency of the continuous wavelet transform of the noise signal y(t);

[0035] Step 3.3, the continuous wavelet coefficients of the noise signal y(t) are obtained by continuous wavelet transform through equation (9):

[0036]

[0037] Among them, K y The continuous wavelet coefficients of the noise signal y(t) after continuous wavelet transform; b is the translation parameter; is the wavelet function, and the complex Morlet wavelet function is used. The formula is as follows:

[0038]

[0039] Among them, F b is the bandwidth parameter; π is the circumference of a circle; e is a natural constant; i is an imaginary unit;

[0040] Step 3.4, using the actual frequency F y , continuous wavelet coefficient Ky and sampling time t to draw the time-frequency diagram of the noise signal p y .

[0041] The specific process of step 4 is:

[0042] Step 4.1, calculate the time-frequency diagram p of the vibration displacement signal x The maximum difference between the corresponding pixels of the fixed window and the sliding window In the time-frequency diagram of vibration displacement signal p x At point (i, j) on the i,j , covering the time-frequency diagram p x m×m pixels around the position (i, j); the other is set up in the time-frequency graph p x At (i', j'), there is a submatrix of size m×m, which also covers the m×m pixels around (i', j'), called the sliding window y i',j' ; Define the maximum difference between the corresponding pixels of the two windows The calculation method is as follows:

[0043]

[0044] Where k = l = m-1 and (i, j) ≠ (i', j'), m is the embedding dimension;

[0045] Step 4.2, calculate the time-frequency diagram p of the vibration displacement signal when the embedding dimension m = 1 and m = 2 respectively x The average value of the number of similarities between the sliding window and the fixed window C x1 and C x2 , the calculation process is as follows:

[0046]

[0047] D1=(max(i)-1)×(max(j)-1) (14)

[0048] D2=(max(i)-2)×(max(j)-2) (15)

[0049]

[0050] Among them, θ is the step function, r is the tolerance parameter to measure whether the two types of windows are similar;

[0051] Step 4.3, calculate the time-frequency diagram p of the noise signal y The maximum difference between the corresponding pixels of the fixed window and the sliding window In the time-frequency diagram of the noise signal p yAt point (i, j) on the i,j , covering the time-frequency diagram p y m×m pixels around the position (i, j); the other is set up in the time-frequency graph p y At (i', j'), there is a submatrix of size m×m, which also covers the m×m pixels around (i', j') and is called the sliding window y i',j' ; The similarity between two windows is determined by the distance between the two windows Decided, It is defined as the maximum difference between the corresponding pixels of the two windows. The calculation process is as follows:

[0052]

[0053] Where k = l = m-1 and (i, j) ≠ (i', j'), m is the embedding dimension;

[0054] Step 4.4, calculate the time-frequency diagram p of the noise signal when the embedding dimension m = 1 and m = 2 respectively y The average values ​​of the number of similarities between the sliding window and the fixed window, Cy1 and Cy2, are calculated as follows:

[0055]

[0056] D1=(max(i)-1)×(max(j)-1) (20)

[0057] D2=(max(i)-2)×(max(j)-2) (21)

[0058]

[0059] The specific process of step 5 is:

[0060] The fusion characteristic value S of the integrated spindle vibration displacement signal and noise signal is calculated by equations (23) to (25): x,y , the calculation process is as follows:

[0061]

[0062] in, and They represent the eigenvalues ​​when the embedding dimension m is 2 and 1 respectively.

[0063] The beneficial effect of the present invention is that the present invention improves the accuracy and reliability of mixed flow pump vibration detection based on the time domain signals of two measuring points through a mixed flow pump stability evaluation method based on acoustic-vibration fusion eigenvalues. The mixed flow pump stability evaluation method based on acoustic-vibration fusion eigenvalues ​​is used to quantitatively evaluate the operating stability of the mixed flow pump, which provides a reliable diagnostic basis and helps to timely discover potential problems and perform maintenance and optimization. The mixed flow pump stability evaluation method based on acoustic-vibration fusion eigenvalues ​​is not only suitable for vibration detection of mixed flow pump systems, but can also be extended to signal analysis of other complex mechanical systems. It has strong versatility and practicality, thereby improving the operating stability, safety and service life of the equipment and reducing maintenance costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 It is a structural schematic diagram of a mixed flow pump vibration and noise data acquisition test system used in the mixed flow pump stability evaluation method based on acoustic-vibration fusion characteristic values ​​of the present invention;

[0065] FIG2( a ) is a diagram of the main shaft vibration displacement signal collected by a laser vibrometer in working condition 1 in the mixed flow pump stability assessment method based on acoustic-vibration fusion characteristic values ​​of the present invention;

[0066] FIG2( b ) is a diagram of the noise signal between the impeller and the guide vane in working condition 1 collected by the noise sensor in the mixed flow pump stability assessment method based on the acoustic-vibration fusion characteristic value of the present invention;

[0067] FIG2(c) is a diagram of the main shaft vibration displacement signal collected by a laser vibrometer under working condition 2 in the mixed flow pump stability assessment method based on acoustic-vibration fusion characteristic values ​​of the present invention;

[0068] FIG2(d) is a diagram of the noise signal between the impeller and the guide vane in working condition 2 collected by the noise sensor in the mixed flow pump stability assessment method based on the acoustic-vibration fusion characteristic value of the present invention;

[0069] FIG3( a ) is a time-frequency image of the main shaft vibration signal used to extract the fusion eigenvalue of working condition 1 in the mixed flow pump stability assessment method based on acoustic-vibration fusion eigenvalue of the present invention;

[0070] FIG3( b ) is a time-frequency image of a noise signal used to extract a fusion characteristic value of working condition 1 in the mixed flow pump stability assessment method based on acoustic-vibration fusion characteristic values ​​of the present invention;

[0071] FIG3( c ) is a time-frequency image of the main shaft vibration signal used to extract the fusion eigenvalue of working condition 2 in the mixed flow pump stability assessment method based on acoustic-vibration fusion eigenvalue of the present invention;

[0072] FIG3( d ) is a time-frequency image of a noise signal used to extract a fusion characteristic value of working condition 2 in the mixed flow pump stability assessment method based on acoustic-vibration fusion characteristic values ​​of the present invention;

[0073] Figure 4 It is the comparative analysis result of the fusion characteristic values ​​of working condition 1 and working condition 2 in the mixed flow pump stability evaluation method based on the acoustic-vibration fusion characteristic value of the present invention.

[0074] In the figure, 1. inlet flow channel, 2. impeller, 3. guide vane, 4. outlet flow channel, 5. laser vibrometer, 6. noise sensor, 7. main shaft, 8. drive motor, 9. data acquisition card, 10. computer. DETAILED DESCRIPTION

[0075] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0076] Example 1

[0077] The mixed flow pump stability evaluation method based on acoustic-vibration fusion characteristic value of the present invention is as follows: Figure 1 As shown, in the mixed flow pump test system, the real-time vibration signal and noise signal are collected and analyzed respectively. The mixed flow pump is controlled by a driving motor 8, and the system includes: an inlet flow channel 1, an impeller 2, a guide vane 3, and an outlet flow channel 4 in sequence according to the water flow direction. For any mixed flow pump device including an inlet flow channel 1, an impeller 2, a guide vane 3, and an outlet flow channel 4, a visualization window is opened near the shaft head of the mixed flow pump for the laser vibrometer 5 to measure the vibration of the main shaft 7. A noise measurement point is set between the impeller and the guide vane of the mixed flow pump, and the noise generated during the operation of the mixed flow pump is measured by a noise sensor 6. The laser vibrometer 5 and the noise sensor 6 are both connected to a data acquisition card 9 through a data transmission line, and transmitted to a computer 10 through a data line.

[0078] Example 2

[0079] The mixed flow pump stability assessment method based on acoustic-vibration fusion characteristic values ​​of the present invention is specifically implemented according to the following steps:

[0080] Working conditions description:

[0081] The working condition 1 of the present invention is that the speed of the mixed flow pump is started in a convex function manner, and the function expression is as follows:

[0082]

[0083] Working condition 2 is that the speed of the mixed flow pump is started in a concave function manner, and the function expression is as follows:

[0084]

[0085] Where n t is the speed of the mixed flow pump at different times, in r / min; n r is the target speed for starting, which is 1800r / min; T Fis the total duration of the startup process, which is 5s; t is any time during the startup process, in seconds.

[0086] Step 1: Start the mixed flow pump vibration and noise data acquisition system. Figure 1 As shown, it includes a Doppler laser vibrometer 5, which is connected to a data acquisition card 9, which is respectively connected to a computer 10 and a noise sensor 6, and the water flows through the inlet flow channel 1, the impeller 2, the guide vane 3 and the outlet flow channel 4 in sequence. The Doppler laser vibrometer 5 is used to collect the vibration displacement signal of the main shaft 7, and the noise sensor 6 is used to collect the noise signal between the impeller 2 and the guide vane 3. The collected signals are collected by the data acquisition card 9 and transmitted to the computer 10 through the data line, and the vibration displacement signal x(t) component diagram shown in Figure 2(a) and Figure 2(c) and the noise signal y(t) component diagram shown in Figure 2(b) and Figure 2(d) are obtained, and t is the sampling time.

[0087] Step 2: Use continuous wavelet transform to convert the vibration signal x(t) collected by the laser vibrometer as shown in Figure 2(a) (c) into the time-frequency diagram p as shown in Figure 3(a) and Figure 3(c) x The specific steps are as follows:

[0088] Step 2.1, calculate the scale parameter of the continuous wavelet transform of the vibration signal x(t). The calculation method is as follows

[0089] a x =2·F xc ·T (3)

[0090] Among them, a x is the scale parameter; F xc is the center frequency of the wavelet function, which is 3; T is the total scale of the continuous wavelet transform of the vibration signal x(t), and its calculation formula is:

[0091]

[0092] In the formula, f s The sampling frequency of the data acquisition card of this experimental device is 10.24kHz;

[0093] Step 2.2: Set the scale parameter a x Convert to actual frequency. The calculation method is as follows:

[0094]

[0095] In the formula, F x is the actual frequency of the continuous wavelet transform of the vibration signal x(t);

[0096] Step 2.3, the continuous wavelet coefficients of the vibration signal x(t) are obtained by continuous wavelet transform through equation (4):

[0097]

[0098] Among them, K x is the continuous wavelet coefficient of the continuous wavelet transform of the vibration signal x(t); b is the translation parameter, ranging from 0 to t, with a step size of is a wavelet function. In the mixed flow pump stability assessment method based on the acoustic-vibration fusion eigenvalue of the present invention, a complex Morlet wavelet function is used, and its formula is as follows:

[0099]

[0100] Among them, F b is the bandwidth parameter, its value is 3; π is the circumference of a circle; e is a natural constant; i is an imaginary unit;

[0101] Step 2.4, using the actual frequency F obtained above x , continuous wavelet coefficient K x The time-frequency diagrams of the vibration displacement signal p are plotted at the sampling time t as shown in Figure 3(a) and Figure 3(c). x .

[0102] Step 3: Use continuous wavelet transform to convert the noise signal y(t) collected by the noise sensor as shown in Figure 2(b) and Figure 2(d) into the time-frequency graph p as shown in Figure 3(b) and Figure 3(d) y The specific steps are as follows:

[0103] Step 3.1, calculate the scale parameter of the continuous wavelet transform of the noise signal y(t). The calculation method is as follows

[0104] a y =2·F yc ·T (8)

[0105] Among them, a y is the scale parameter; F yc is the center frequency of the wavelet function, which is 3; T is the total scale of the continuous wavelet transform of the noise signal y(t), and its calculation formula is:

[0106]

[0107] In the formula, f s The sampling frequency of the data acquisition card of this experimental device is 10.24kHz;

[0108] Step 3.2: Set the scale parameter a y Convert to actual frequency. The calculation method is as follows:

[0109]

[0110] In the formula, F y is the actual frequency of the continuous wavelet transform of the noise signal y(t);

[0111] Step 3.3, the continuous wavelet coefficients of the noise signal y(t) are obtained by continuous wavelet transform through equation (11):

[0112]

[0113] Among them, K y The continuous wavelet coefficients of the noise signal y(t) are transformed by continuous wavelet transform; b is the translation parameter, ranging from 0 to t, with a step size of is a wavelet function. In the mixed flow pump stability assessment method based on the acoustic-vibration fusion eigenvalue of the present invention, a complex Morlet wavelet function is used, and its formula is as follows:

[0114]

[0115] Among them, F b is the bandwidth parameter, its value is 3; π is the circumference of a circle; e is a natural constant; i is an imaginary unit;

[0116] Step 3.4, using the actual frequency F obtained above y , continuous wavelet coefficient K y The noise signal time-frequency diagram p is plotted at the sampling time t as shown in Figure 3(b)(d). y .

[0117] Step 4: Based on the result of step 3, the time-frequency diagram p of the vibration displacement signal as shown in FIG3(a) and FIG3(c) is calculated by the following method: x And the time-frequency diagram of the noise signal as shown in Figure 3(b) and Figure 3(d) y The average value of the number of similarities between the sliding window and the fixed window C x and C y The specific calculation steps are as follows:

[0118] Step 4.1, calculate the time-frequency diagram p of the vibration displacement signal as shown in Figure 3(a) and Figure 3(c) x The maximum difference between the corresponding pixels of the fixed window and the sliding window In the time-frequency diagram of vibration displacement signal p x At point (i, j) on the i,j , which covers the time-frequency diagram p xm×m pixels around the position (i, j); the other is set up in the time-frequency graph p x At (i', j'), there is a submatrix of size m×m. This submatrix also covers the m×m pixels around (i', j') and can slide on the image to cover different areas. This is called the sliding window y i',j' ; The similarity between two windows is determined by the distance between the two windows Decided, It is defined as the maximum difference between the corresponding pixels of the two windows, which is calculated as follows:

[0119]

[0120] Where k = l = m-1 and (i, j) ≠ (i', j') to avoid comparing a fixed window with itself, and m is the embedding dimension;

[0121] Step 4.2, calculate the time-frequency diagram p of the vibration displacement signal when the embedding dimension m = 1 and m = 2, as shown in Figure 3 (a) (c) x The average value of the number of similarities between the sliding window and the fixed window C x1 and C x2 , the calculation method is as follows:

[0122]

[0123] D1=(max(i)-1)×(max(j)-1) (16)

[0124] D2=(max(i)-2)×(max(j)-2) (17)

[0125]

[0126] Among them, θ is a step function, r is a tolerance parameter to measure whether two types of windows are similar. If the maximum difference in pixels between the two windows is If it is less than the tolerance parameter r, the windows are considered similar, and the r value ranges from 0.26 to 0.6 with a step size of 0.02;

[0127] Step 4.3, calculate the time-frequency diagram p of the noise signal as shown in Figure 3(b) and Figure 3(d) y The maximum difference between the corresponding pixels of the fixed window and the sliding window In the time-frequency diagram of the noise signal p y At point (i, j) on the i,j , which covers the time-frequency diagram p y m×m pixels around the position (i, j); the other is set up in the time-frequency graph p yAt (i', j'), there is a submatrix of size m×m. This submatrix also covers the m×m pixels around (i', j') and can slide on the image to cover different areas. This is called the sliding window y i',j' ; The similarity between two windows is determined by the distance between the two windows Decided, It is defined as the maximum difference between the corresponding pixels of the two windows, which is calculated as follows:

[0128]

[0129] Where k = l = m-1 and (i, j) ≠ (i', j') to avoid comparing a fixed window with itself, and m is the embedding dimension;

[0130] Step 4.4, calculate the time-frequency diagram p of the noise signal when the embedding dimension m = 1 and m = 2, as shown in Figure 3(d) and Figure 3(d) respectively. y The average values ​​of the number of similarities between the sliding window and the fixed window, Cy1 and Cy2, are calculated as follows:

[0131]

[0132] D1=(max(i)-1)×(max(j)-1) (22)

[0133] D2=(max(i)-2)×(max(j)-2) (23)

[0134]

[0135] Among them, θ is a step function, r is a tolerance parameter to measure whether two types of windows are similar. If the maximum difference in pixels between the two windows is If it is less than the tolerance parameter r, the windows are considered similar, and the r value ranges from 0.26 to 0.6 with a step size of 0.02;

[0136] Step 5: Based on step 4, we combine different signals and C under different embedding dimensions. x and C y Build C x,y , the fusion characteristic value S of the integrated spindle vibration displacement signal and noise signal is calculated by equations (25) to (27) x,y ; The calculation method is:

[0137]

[0138]

[0139] in, and They represent the eigenvalues ​​when the embedding dimension m is 2 and 1 respectively.

[0140] Example 3

[0141] Obtain the fusion eigenvalue S that comprehensively considers the vibration displacement signal and the noise signal x,y Then, draw Figure 4 The fusion eigenvalue S x,y —Tolerance parameter r graph, S x,y The larger the value, the more complex the dynamic characteristics of the signal in the time-frequency domain. Therefore, the fusion eigenvalue S of the mixed flow pump acoustic vibration signal is x,y The larger the value, the worse the pump stability.

Claims

1. A mixed flow pump stability assessment method based on acoustic-vibration fusion eigenvalues, characterized by: The specific steps include: Step 1, collecting the spindle vibration displacement signal and noise signal; Step 2: Use continuous wavelet transform to convert the spindle vibration displacement signal collected in step 1 into a time-frequency graph p x ; Step 3: Use continuous wavelet transform to convert the noise signal collected in step 1 into a time-frequency graph p y ; Step 4: Obtain the time-frequency graph p x And the time-frequency diagram p y The average value of the number of similarities between the sliding window and the fixed window C x and C y ; Step 5: By combining different signals and C under different embedding dimensions x and C y Build C x,y , calculate the fusion eigenvalue S of the integrated spindle vibration displacement signal and noise signal x,y .

2. The mixed flow pump stability assessment method based on acoustic-vibration fusion characteristic value according to claim 1 is characterized in that: The specific process of step 1 is as follows: the water flows through the inlet flow channel (1), the impeller (2), the guide vane (3) and the outlet flow channel (4) in sequence, a Doppler laser vibrometer (5) is used to collect the vibration displacement signal of the main shaft (7), and a noise sensor (6) is used to collect the noise signal between the impeller (2) and the guide vane (3), the collected signals are collected by a data acquisition card (9), and transmitted to a computer (10) through a data line, to obtain a vibration displacement signal x(t) component diagram and a noise signal y(t) component diagram, where t is the sampling time.

3. The mixed flow pump stability assessment method based on acoustic-vibration fusion characteristic value according to claim 2 is characterized in that: The specific process of step 2 is: Step 2.1, use the following formula (1) to calculate the scale parameter of the continuous wavelet transform of the vibration signal x(t): a x =2·F xc ·T (1) Among them, a x is the scale parameter; F xc is the center frequency of the wavelet function; T is the total scale of the continuous wavelet transform of the vibration signal x(t), which is calculated using the following formula (2): In the formula, f s The sampling frequency of the data acquisition card of this experimental device; Step 2.2: Use the following formula (3) to convert the scale parameter a x Convert to actual frequency: In the formula, F x is the actual frequency of the continuous wavelet transform of the vibration signal x(t); Step 2.3, the continuous wavelet coefficients of the vibration signal x(t) are obtained by continuous wavelet transform through equation (4): Among them, K x is the continuous wavelet coefficient of the continuous wavelet transform of the vibration signal x(t); b is the translation parameter, is the wavelet function, and the complex Morlet wavelet function is used. The formula is as follows: Among them, F b is the bandwidth parameter; π is the circumference of a circle; e is a natural constant; i is an imaginary unit; Step 2.4, using the actual frequency F x , continuous wavelet coefficient K x and sampling time t to draw the time-frequency diagram of the vibration displacement signal p x .

4. The mixed flow pump stability assessment method based on acoustic-vibration fusion characteristic value according to claim 3 is characterized in that: The specific process of step 3 is as follows: Step 3.1, use the following formula (6) to calculate the scale parameter of the continuous wavelet transform of the noise signal y(t): a y =2·F yc ·T (6) Among them, a y is the scale parameter; F yc is the center frequency of the wavelet function; T is the total scale of the noise signal y(t) for continuous wavelet transform, and the calculation formula is: In the formula, f s The sampling frequency of the data acquisition card of this experimental device; Step 3.2: Use the following formula (8) to convert the scale parameter a y Convert to actual frequency: In the formula, F y is the actual frequency of the continuous wavelet transform of the noise signal y(t); Step 3.3, the continuous wavelet coefficients of the noise signal y(t) are obtained by continuous wavelet transform through equation (9): Among them, K y The continuous wavelet coefficients of the noise signal y(t) after continuous wavelet transform; b is the translation parameter; is the wavelet function, and the complex Morlet wavelet function is used. The formula is as follows: Among them, F b is the bandwidth parameter; π is the circumference of a circle; e is a natural constant; i is an imaginary unit; Step 3.4, using the actual frequency F y , continuous wavelet coefficient K y and sampling time t to draw the time-frequency diagram of the noise signal p y .

5. The mixed flow pump stability assessment method based on acoustic-vibration fusion characteristic value according to claim 4 is characterized in that: The specific process of step 4 is as follows: Step 4.1, calculate the time-frequency diagram p of the vibration displacement signal x The maximum difference between the corresponding pixels of the fixed window and the sliding window In the time-frequency diagram of vibration displacement signal p x At point (i, j) on the i,j , covering the time-frequency diagram p x m×m pixels around the position (i, j); the other is set up in the time-frequency graph p x At (i', j'), there is a submatrix of size m×m, which also covers the m×m pixels around (i', j'), called the sliding window y i',j' ; Define the maximum difference between the corresponding pixels of the two windows The calculation method is as follows: Where k = l = m-1 and (i, j) ≠ (i', j'), m is the embedding dimension; Step 4.2, calculate the time-frequency diagram p of the vibration displacement signal when the embedding dimension m = 1 and m = 2 respectively x The average value of the number of similarities between the sliding window and the fixed window C x1 and C x2 , the calculation process is as follows: D1=(max(i)-1)×(max(j)-1) (14) D2=(max(i)-2)×(max(j)-2) (15) Among them, θ is the step function, r is the tolerance parameter to measure whether the two types of windows are similar; Step 4.3, calculate the time-frequency diagram p of the noise signal y The maximum difference between the corresponding pixels of the fixed window and the sliding window In the time-frequency diagram of the noise signal p y At point (i, j) on the i,j , covering the time-frequency diagram p y m×m pixels around the position (i, j); the other is set up in the time-frequency graph p y At (i', j'), there is a submatrix of size m×m, which also covers the m×m pixels around (i', j') and is called the sliding window y i',j' ; The similarity between two windows is determined by the distance between the two windows Decided, It is defined as the maximum difference between the corresponding pixels of the two windows. The calculation process is as follows: Where k = l = m-1 and (i, j) ≠ (i', j'), m is the embedding dimension; Step 4.4, calculate the time-frequency diagram p of the noise signal when the embedding dimension m = 1 and m = 2 respectively y The average values ​​of the number of similarities between the sliding window and the fixed window, Cy1 and Cy2, are calculated as follows: D1=(max(i)-1)×(max(j)-1) (20) D2=(max(i)-2)×(max(j)-2) (21) 6. The mixed flow pump stability assessment method based on acoustic-vibration fusion characteristic value according to claim 5 is characterized in that: The specific process of step 5 is as follows: The fusion characteristic value S of the integrated spindle vibration displacement signal and noise signal is calculated by equations (23) to (25): x,y , the calculation process is as follows: in, and They represent the eigenvalues ​​when the embedding dimension m is 2 and 1 respectively.

Images